Comment on “Data-driven Subtypes of Polycystic Ovary Syndrome and Their Association With Clinical Outcomes”

An estimated 11% to 13% of women worldwide are affected by polycystic ovary syndrome (PCOS), and it has profound impacts both short and long-term on reproduction and metabolism. PCOS is associated with health conditions such as obesity, type 2 diabetes mellitus, insulin resistance, and cardiovascular disease, and can have a significant impact on quality of life. PCOS can present with a wide range of symptoms, which presents some difficulty in both diagnosis and treatment of the condition. There is a lack of evidence and clinical agreement on subtypes of PCOS; this study presents 4 distinct subtypes of PCOS that were validated in 5 cohorts of ethnically diverse individuals and assesses the association between subtypes and in vitro fertilization (IVF) outcomes and pregnancy complications over a 6.5-year follow-up period. The cohort used to define the subtypes included 11,908 women with PCOS who were not receiving therapy at baseline. Initially, 29 clinically relevant variables were included, which were narrowed to 9 for clustering. The 4 subtypes of PCOS identified were hyperandrogenic (HA-PCOS), obesity (OB-PCOS), high-sex hormone-binding globulin (SHBG-PCOS), and high luteinizing hormone and anti-Müllerian hormone (LH-PCOS). Distinct clinical characteristics included mild metabolic disorders and high testosterone-dehydroepiandrosterone sulfate (DHEA-S) in HA-PCOS, higher body mass index, fasting glucose, and fasting insulin, as well as high prevalence of type 2 diabetes mellitus in OB-PCOS, high sex hormone-binding globulin and lowest BMI in SHBG-PCOS, and elevated luteinizing, anti-Müllerian, and follicle-stimulating hormones in LH-PCOS. Results were replicated in 5 validation cohorts from China (1476 patients), the United States (593 patients), Europe (197 patients), Singapore (127 patients), and Brazil (85 patients). Longitudinal follow-up included 4542 patients between 2014 and 2018. Follow-up evaluation included PCOS remission, hyperandrogenic, ovulatory and polycystic ovarian conditions, and cumulative incidence of chronic metabolic complications. The proportion of patients still meeting Rotterdam criteria at follow-up was highest in the LH-PCOS group, followed by HA-PCOS, SHBG-PCOS, and OB-PCOS. Persistent hyperandrogenic state was highest in the HA- and LH-PCOS groups, followed by the SHBG and OB-PCOS groups. SHBG-PCOS also had the lowest rates of ovulatory dysfunction and polycystic ovaries at follow-up. In terms of IVF outcomes, the highest live birth rate was experienced in the SHBG-PCOS group, followed by LH-PCOS, HA-PCOS, and OB-PCOS groups. Both SHBG and LH-PCOS groups had higher live birth rates than controls. Clinical pregnancy rates for all groups were higher than the control group, as were pregnancy loss rates. The highest pregnancy loss rate was observed in HA-PCOS, and OB-PCOS had the poorest outcomes overall. All PCOS groups had an increased risk of moderate to severe ovarian hyperstimulation syndrome (OHSS), and some groups showed a higher risk of gestational diabetes and gestational hypertension. These results indicate that these 4 subtypes of PCOS are robust and clinically meaningful in assisting with the prediction of outcomes in women with PCOS. Future research should focus on further definition of these subtypes and their association with outcomes, as well as potentially targeted interventions meant to optimize outcomes for specific subtypes of PCOS. In addition, continuous measures of ovarian and menstrual cycle dysfunction should be explored, and biomarkers should be identified to assist in the diagnosis and treatment of PCOS. (Summarized from Gao X, Zhao S, Du Y, et al. Data-driven subtypes of polycystic ovary syndrome and their association with clinical outcomes. Nat Med . 2025;31(12):4214-4224. doi: 10.1038/s41591-025-03984-1)